OptiDistillNet: Learning nonlinear pulse propagation using the student-teacher model.

We present a unique approach for learning the pulse evolution in a nonlinear fiber using a deep convolutional neural network (CNN) by solving the nonlinear Schrodinger equation (NLSE). Deep network model compression has become widespread for deploying such models in real-world applications. A knowledge distillation (KD) based framework for compressing a CNN is presented here. The student network, termed here as OptiDistillNet has better generalisation, has faster convergence, is faster and uses less number of trainable parameters. This work represents the first effort, to the best of our knowledge, that successfully applies a KD-based technique for any nonlinear optics application. Our tests show that even by reducing the model size by up to 91.2%, we can still achieve a mean square error (MSE) which is very close to the MSE of 1.04*10-5 achieved by the teacher model. The advantages of the suggested model include the use of a simple architecture, fast optimization, and improved accuracy, opening up applications in optical coherent communication systems.

Modeling and prediction of COVID-19 pandemic using Gaussian mixture model.

COVID-19 is caused by a novel coronavirus and has played havoc on many countries across the globe. A majority of the world population is now living in a restricted environment for more than a month with minimal economic activities, to prevent exposure to this highly infectious disease. Medical professionals are going through a stressful period while trying to save the larger population. In this paper, we develop two different models to capture the trend of a number of cases and also predict the cases in the days to come, so that appropriate preparations can be made to fight this disease. The first one is a mathematical model accounting for various parameters relating to the spread of the virus, while the second one is a non-parametric model based on the Fourier decomposition method (FDM), fitted on the available data. The study is performed for various countries, but detailed results are provided for the India, Italy, and United States of America (USA). The turnaround dates for the trend of infected cases are estimated. The end-dates are also predicted and are found to agree well with a very popular study based on the classic susceptible-infected-recovered (SIR) model. Worldwide, the total number of expected cases and deaths are 12.7 × 106 and 5.27 × 105, respectively, predicted with data as of 06-06-2020 and 95% confidence intervals. The proposed study produces promising results with the potential to serve as a good complement to existing methods for continuous predictive monitoring of the COVID-19 pandemic.

Satellite hyperspectral imaging technology as a potential rapid pollution assessment tool for urban landfill sites: case study of Ghazipur and Okhla landfill sites in Delhi, India.

Hyperspectral imaging technology has been used for biochemical analysis of Earth's surface exploiting the spectral reflectance signatures of various materials. The new-generation Italian PRISMA (PRecursore IperSpettrale dellaMissione Applicativa) hyperspectral satellite launched by the Italian space agency (ASI) provides a unique opportunity to map various materials through spectral signature analysis for recourse management and sustainable development. In this study PRISMA hyperspectral satellite imagery-based multiple spectral indices were generated for rapid pollution assessment at Ghazipur and Okhla landfill sites in Delhi, India. It was found that the combined risk score for Okhla landfill site was higher than the Ghazipur landfill site. Various manmade materials identified, exploiting the hyperspectral imagery and spectral signature libraries, indicated presence of highly saline water, plastic (black, ABS, pipe, netting, etc.), asphalt tar, black tar paper, kerogen BK-Cornell, black paint and graphite, chalcocite minerals, etc. in large quantities in both the landfill sites. The methodology provides a rapid pollution assessment tool for municipal landfill sites.

Robust Prototypical Few-Shot Organ Segmentation With Regularized Neural-ODEs.

Despite the tremendous progress made by deep learning models in image semantic segmentation, they typically require large annotated examples, and increasing attention is being diverted to problem settings like Few-Shot Learning (FSL) where only a small amount of annotation is needed for generalisation to novel classes. This is especially seen in medical domains where dense pixel-level annotations are expensive to obtain. In this paper, we propose Regularized Prototypical Neural Ordinary Differential Equation (R-PNODE), a method that leverages intrinsic properties of Neural-ODEs, assisted and enhanced by additional cluster and consistency losses to perform Few-Shot Segmentation (FSS) of organs. R-PNODE constrains support and query features from the same classes to lie closer in the representation space thereby improving the performance over the existing Convolutional Neural Network (CNN) based FSS methods. We further demonstrate that while many existing Deep CNN-based methods tend to be extremely vulnerable to adversarial attacks, R-PNODE exhibits increased adversarial robustness for a wide array of these attacks. We experiment with three publicly available multi-organ segmentation datasets in both in-domain and cross-domain FSS settings to demonstrate the efficacy of our method. In addition, we perform experiments with seven commonly used adversarial attacks in various settings to demonstrate R-PNODE's robustness. R-PNODE outperforms the baselines for FSS by significant margins and also shows superior performance for a wide array of attacks varying in intensity and design.

Review on machine learning-based bioprocess optimization, monitoring, and control systems.

Machine Learning is quickly becoming an impending game changer for transforming big data thrust from the bioprocessing industry into actionable output. However, the complex data set from bioprocess, lagging cyber-integrated sensor system, and issues with storage scalability limit machine learning real-time application. Hence, it is imperative to know the state of technology to address prevailing issues. This review first gives an insight into the basic understanding of the machine learning domain and discusses its complexities for more comprehensive applications. Followed by an outline of how relevant machine learning models are for statistical and logical analysis of the enormous datasets generated to control bioprocess operations. Then this review critically discusses the current knowledge, its limitations, and future aspects in different subfields of the bioprocessing industry. Further, this review discusses the prospects of adopting a hybrid method to dovetail different modeling strategies, cyber-networking, and integrated sensors to develop new digital biotechnologies.

Semantic segmentation of bone structures in chest X-rays including unhealthy radiographs: A robust and accurate approach.

The chest X-ray is a widely used medical imaging technique for the diagnosis of several lung diseases. Some nodules or other pathologies present in the lungs are difficult to visualize on chest X-rays because they are obscured byoverlying boneshadows. Segmentation of bone structures and suppressing them assist medical professionals in reliable diagnosis and organ morphometry. But segmentation of bone structures is challenging due to fuzzy boundaries of organs and inconsistent shape and size of organs due to health issues, age, and gender. The existing bone segmentation methods do not report their performance on abnormal chest X-rays, where it is even more critical to segment the bones. This work presents a robust encoder-decoder network for semantic segmentation of bone structures on normal as well as abnormal chest X-rays. The novelty here lies in combining techniques from two existing networks (Deeplabv3+ and U-net) to achieve robust and superior performance. The fully connected layers of the pre-trained ResNet50 network have been replaced by an Atrous spatial pyramid pooling block for improving the quality of the embedding in the encoder module. The decoder module includes four times upsampling blocks to connect both low-level and high-level features information enabling us to retain both the edges and detail information of the objects. At each level, the up-sampled decoder features are concatenated with the encoder features at a similar level and further fine-tuned to refine the segmentation output. We construct a diverse chest X-ray dataset with ground truth binary masks of anterior ribs, posterior ribs, and clavicle bone for experimentation. The dataset includes 100 samples of chest X-rays belonging to healthy and confirmed patients of lung diseases to maintain the diversity and test the robustness of our method. We test our method using multiple standard metrics and experimental results indicate an excellent performance on both normal and abnormal chest X-rays.

Deep learning-based approach for identification of diseases of maize crop.

In recent years, deep learning techniques have shown impressive performance in the field of identification of diseases of crops using digital images. In this work, a deep learning approach for identification of in-field diseased images of maize crop has been proposed. The images were captured from experimental fields of ICAR-IIMR, Ludhiana, India, targeted to three important diseases viz. Maydis Leaf Blight, Turcicum Leaf Blight and Banded Leaf and Sheath Blight in a non-destructive manner with varied backgrounds using digital cameras and smartphones. In order to solve the problem of class imbalance, artificial images were generated by rotation enhancement and brightness enhancement methods. In this study, three different architectures based on the framework of 'Inception-v3' network were trained with the collected diseased images of maize using baseline training approach. The best-performed model achieved an overall classification accuracy of 95.99% with average recall of 95.96% on the separate test dataset. Furthermore, we compared the performance of the best-performing model with some pre-trained state-of-the-art models and presented the comparative results in this manuscript. The results reported that best-performing model performed quite better than the pre-trained models. This demonstrates the applicability of baseline training approach of the proposed model for better feature extraction and learning. Overall performance analysis suggested that the best-performed model is efficient in recognizing diseases of maize from in-field images even with varied backgrounds.

Riemannian Curvature of Deep Neural Networks.

We analyze deep neural networks using the theory of Riemannian geometry and curvature. The objective is to gain insight into how Riemannian geometry can characterize and predict the trained behavior of neural networks. We define a method for calculating Riemann and Ricci curvature tensors, and Ricci scalar curvature values for a trained neural net, in such a way that the output classifier softmax values are related to the input transformations, through the curvature equations. We also measure these curvature tensors experimentally for different networks which are pretrained with stochastic gradient descent and offer a way of visualizing and understanding the measurements to gain insight into the effect curvature has on behavior the neural networks locally, and possibly predict their behavior for different transformations of the test data. We also analyze the effect of variation in depth of the neural networks as well as how it behaves for different choices of data set.

Tractography-Based Score for Learning Effective Connectivity From Multimodal Imaging Data Using Dynamic Bayesian Networks.

OBJECTIVE: Effective connectivity (EC) is the methodology for determining functional-integration among the functionally active segregated regions of the brain. By definition EC is "the causal influence exerted by one neuronal group on another" which is constrained by anatomical connectivity (AC) (axonal connections). AC is necessary for EC but does not fully determine it, because synaptic communication occurs dynamically in a context-dependent fashion. Although there is a vast emerging evidence of structure-function relationship using multimodal imaging studies, till date only a few studies have done joint modeling of the two modalities: functional MRI (fMRI) and diffusion tensor imaging (DTI). We aim to propose a unified probabilistic framework that combines information from both sources to learn EC using dynamic Bayesian networks (DBNs). METHOD: DBNs are probabilistic graphical temporal models that learn EC in an exploratory fashion. Specifically, we propose a novel anatomically informed (AI) score that evaluates fitness of a given connectivity structure to both DTI and fMRI data simultaneously. The AI score is employed in structure learning of DBN given the data. RESULTS: Experiments with synthetic-data demonstrate the face validity of structure learning with our AI score over anatomically uninformed counterpart. Moreover, real-data results are cross-validated by performing classification-experiments. CONCLUSION: EC inferred on real fMRI-DTI datasets is found to be consistent with previous literature and show promising results in light of the AC present as compared to other classically used techniques such as Granger-causality. SIGNIFICANCE: Multimodal analyses provide a more reliable basis for differentiating brain under abnormal/diseased conditions than the single modality analysis.

Low-cost multifrequency electrical impedance-based system (MFEIBS) for clinical imaging: design and performance evaluation.

Electrical impedance tomography (EIT) is an upcoming and capable imaging modality used for clinical imaging. It is non-invasive, non-ionising and an inexpensive technique. This paper explains the designing and the analysis of a low-cost multifrequency electrical impedance-based system (MFEIBS) having a flexible mechanism of interfacing up to 32 electrodes, suitable for 1 kHz-2 MHz. Various indicators to check the performance of the EIT system were evaluated and presented here. The performance of VCO and VCCS was measured up to 2 MHz. SNR was measured with saline phantom and its mean value is 74 dB for the complete bandwidth. Different combinations of resistors and capacitors were used to find the accuracy of the system, and relative error was less than 0.55% for the entire range. CMRR of the system was calculated and it was found to be maximum 85 dB at 1 kHz frequency. A 16-electrode circular plastic phantom having a diameter of 18 cm was established and connected with a simple MFEIBS. Obtained surface potential was applied to the computer used for image formation using NI USB-6259, 16-bit, 1.25 MS/s M Series High-speed DAQ. Images reconstructed using the system presented in this paper was generated from a 16-electrode plastic phantom filled with NaCl up to 1.2 cm height.

Assessing the correlation of genotypic and phenotypic responses of indica rice varieties under drought stress.

Drought is one of the severe abiotic stress that affects the productivity of rice, an important staple crop that is consumed all over the world. The traits responsible for enhancing or adapting drought resistance in rice plants can be selected and studied to improve their growth under stress conditions. Experiments have been conducted on indica rice varieties comprising Sahabhagidhan as drought tolerant variety and IR64, MTU1010 categorized as drought sensitive varieties. Various root related biochemical and morphological traits such as root length, relative water content (RWC), xylem number, xylem area, proline content, and malondialdehyde content have been investigated for a comparative study of the plant response to drought stress in different rice varieties. The results of differential root transcriptome analysis have revealed that there is a notable difference in gene expression of OsPIP2;5 and OsNIP2;1 in various indica varieties of rice at different time periods of stress. The present work aims at assessing the correlation between genotypic and phenotypic traits that can contribute towards the emerging field of rice phenomics.

The dynamic programming high-order Dynamic Bayesian Networks learning for identifying effective connectivity in human brain from fMRI.

BACKGROUND: Determination of effective connectivity (EC) among brain regions using fMRI is helpful in understanding the underlying neural mechanisms. Dynamic Bayesian Networks (DBNs) are an appropriate class of probabilistic graphical temporal-models that have been used in past to model EC from fMRI, specifically order-one. NEW-METHOD: High-order DBNs (HO-DBNs) have still not been explored for fMRI data. A fundamental problem faced in the structure-learning of HO-DBN is high computational-burden and low accuracy by the existing heuristic search techniques used for EC detection from fMRI. In this paper, we propose using dynamic programming (DP) principle along with integration of properties of scoring-function in a way to reduce search space for structure-learning of HO-DBNs and finally, for identifying EC from fMRI which has not been done yet to the best of our knowledge. The proposed exact search-&-score learning approach HO-DBN-DP is an extension of the technique which was originally devised for learning a BN's structure from static data (Singh and Moore, 2005). RESULTS: The effectiveness in structure-learning is shown on synthetic fMRI dataset. The algorithm reaches globally-optimal solution in appreciably reduced time-complexity than the static counterpart due to integration of properties. The proof of optimality is provided. COMPARISON: The results demonstrate that HO-DBN-DP is comparably more accurate and faster than currently used structure-learning algorithms used for identifying EC from fMRI. The real data EC from HO-DBN-DP shows consistency with previous literature than the classical Granger Causality method. CONCLUSION: Hence, the DP algorithm can be employed for reliable EC estimates from experimental fMRI data.

Learning effective connectivity from fMRI using autoregressive hidden Markov model with missing data.

BACKGROUND: Effective connectivity (EC) analysis of neuronal groups using fMRI delivers insights about functional-integration. However, fMRI signal has low-temporal resolution due to down-sampling and indirectly measures underlying neuronal activity. NEW METHOD: The aim is to address above issues for more reliable EC estimates. This paper proposes use of autoregressive hidden Markov model with missing data (AR-HMM-md) in dynamically multi-linked (DML) framework for learning EC using multiple fMRI time series. In our recent work (Dang et al., 2016), we have shown how AR-HMM-md for modelling single fMRI time series outperforms the existing methods. AR-HMM-md models unobserved neuronal activity and lost data over time as variables and estimates their values by joint optimization given fMRI observation sequence. RESULTS: The effectiveness in learning EC is shown using simulated experiments. Also the effects of sampling and noise are studied on EC. Moreover, classification-experiments are performed for Attention-Deficit/Hyperactivity Disorder subjects and age-matched controls for performance evaluation of real data. Using Bayesian model selection, we see that the proposed model converged to higher log-likelihood and demonstrated that group-classification can be performed with higher cross-validation accuracy of above 94% using distinctive network EC which characterizes patients vs. CONTROLS: The full data EC obtained from DML-AR-HMM-md is more consistent with previous literature than the classical multivariate Granger causality method. COMPARISON: The proposed architecture leads to reliable estimates of EC than the existing latent models. CONCLUSIONS: This framework overcomes the disadvantage of low-temporal resolution and improves cross-validation accuracy significantly due to presence of missing data variables and autoregressive process.

Assessing assumptions of multivariate linear regression framework implemented for directionality analysis of fMRI.

Directionality analysis of time-series, recorded from task-activated regions-of-interest (ROIs) during functional Magnetic Resonance Imaging (fMRI), has helped in gaining insights of complex human behavior and human brain functioning. The most widely used standard method of Granger Causality for evaluating directionality employ linear regression modeling of temporal processes. Such a parameter-driven approach rests on various underlying assumptions about the data. The short-comings can arise when misleading conclusions are reached after exploration of data for which the assumptions are getting violated. In this study, we assess assumptions of Multivariate Autoregressive (MAR) framework which is employed for evaluating directionality among fMRI time-series recorded during a Sensory-Motor (SM) task. The fMRI time-series here is an averaged time-series from a user-defined ROI of multiple voxels. The "aim" is to establish a step-by-step procedure using statistical methods in conjunction with graphical methods to seek the validity of MAR models, specifically in the context of directionality analysis of fMRI data which has not been done previously to the best of our knowledge. Here, in our case of SM task (block design paradigm) there is violation of assumptions, indicating the inadequacy of MAR models to find directional interactions among different task-activated regions of brain.